BYTE.com > ActiveXplained > 2000 > December
Understanding Microsoft .NET, Part I
By David Platt
December 21, 2000
(Understanding Microsoft .NET, Part I
: Page 1 of 4 )
The computer and non-computer press has carried an enormous amount of hype over the past six months, even for this industry, about Microsoft's forthcoming .NET product.
As usual, the marketing blizzard contains essentially no useful information (see, for example, www.microsoft.com/business/vision/netvision.asp, about as enlightening as a Bush-Gore debate on toenail fungus in gerbils).
Still, the notion that something large and profitable is about to happen has taken root in the developer community, demonstrated by the fact that the developers' conference that explained it drew 5,000 attendees to Orlando's punishing heat last July. The first beta version of the software just shipped, so I'm devoting this column and at least my next two columns to unraveling the hype, explaining exactly what .NET is and how you can make money with it.
The PC desktop is a relatively mature market. No one except Microsoft makes any money selling word processors or spreadsheets any more. On the other hand, the Internet is an enormous vacuum, just as the PC desktop was a decade ago. Writing applications to fill it will be highly profitable, but writing good applications for it is extremely difficult. Microsoft's .NET architecture (platform, framework, whatever they're calling it this week) is a development and runtime environment that makes it easier and faster, and hence cheaper, for programmers to develop well-constructed Internet applications. For me, .NET makes the most sense if I think of it in terms of three areas of functionality:
1. A New Object Runtime Environment
The primary drawback of COM has always been its low level of OS support. By this, I mean that a developer wanting to write a COM object has to spend her own resources providing a fairly high level of required infrastructure. For example, every object has to provide its own reference counting and lifetime management code in the form of the AddRef and Release methods of the IUnknown interface, and every object client has to remember to call these at the right times or objects destruct prematurely or get leaked away.
BYTE.com > ActiveXplained > 2000 > December
|
